Segmentation is a critical step in biomedical image analysis, enabling the isolation of individual structures from 3D volumes. However, manual segmentation is time-consuming and limits the scope for comparative research. Automated segmentation using Biomedisa and Deep Learning allows for large-scale analysis of insect brains, providing insights into animal behavior, ecology, and evolution.

In this talk, I will demonstrate the use of Biomedisa for automated segmentation and analysis of 3D image data from 187 honey bee and bumblebee brains. Biomedisa can be used online or locally for semi-automated segmentation to create training data and subsequently train neural networks.

The analysis revealed strong inter-individual variations in total brain size, potentially underpinning behavioural variability central to complex social organisations. Additionally, the bumblebee dataset showed a significant level of lateralization in optic and antennal lobes, which may explain reported variations in visual and olfactory learning.